Sensor Fusion Algorithms For Measuring Multi-DOF Orientation Of Spherical Motors

Spherical motot is a kind of electromagnetic motor which can provide three degree of freedom motion based on a central ball-like joint.It has compact structure,nonsingular moving range and rotate smoothly,omit the need of bulky transmission system compared with the tranditional Multi-DOF platform,which makes it have great potential for the application of robotic joints and conformal printing.In the application with Spherical motor,an attitude measure system with precise result and high data rate is very important for the close-loop control system,but present method such as the coding image or the assist framework have many limits on the data rate or flexibility.Inspired by the attitude fusion system in the aircraft,which combine the measure data of different sensors and compensate the deficiency of different sensors with the fusion algorithm,this paper will research the algorithm of attitude fusion for the Spherical motor in Multi-DOF system.The nomal inertia measure system has to face the problem of geomagnetic disturbance and acceleration disturbance.The key of Multi-DOF attitude fusion algorithm is how to overcome the measure disturbance and provide high quality attitude feedback.This paper begins with the measure model of gyroscope and accelerometer,put out the attidude measures system and the translation between the measural attitude and rotor’s attitude.Base on the EKF method,detailed derivation of attitude fusion algorithm is presented.A method with the measure of acceleration of stator called time-varying reference vector attitude fusion algorithm is proposed for the probem of disturbance in the measure system.Then,a series of simulation data is generated to verify the validity of fusion algorithm.Considering the error of reality,the noise of sensors is measured and analyzed,a suitable filter is design and the calibration method is presented.Finally the algorithm is verified on the experimental platform and the result shows that our fusion algorithm is effective in the orientation compensation and attitude measure.